Skip to content
Closed
Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
112 commits
Select commit Hold shift + click to select a range
74bd3e2
feat: add read-only H3 performance history
edwingao28 Sep 15, 2026
a2142e4
fix: open published H3 results from the video tab
edwingao28 Sep 16, 2026
2ada599
feat: map H3 artifact hardware to HW_REGISTRY keys and TCO tiers
edwingao28 Sep 20, 2026
72669b3
feat: add video benchmark metric registry with TCO tiers and GPU basis
edwingao28 Sep 20, 2026
61a6589
feat: add direction-aware Pareto frontier for video hardware points
edwingao28 Sep 20, 2026
49059f7
feat: add v_ URL state for the video hardware dashboard
edwingao28 Sep 20, 2026
b240682
feat: project GPU counts, power and server layout per published video…
edwingao28 Sep 20, 2026
22187d5
feat: flatten published video history into chart points
edwingao28 Sep 20, 2026
89b247f
test: rebuild video history fixture with per-cell GPU, power and layo…
edwingao28 Sep 20, 2026
2124786
feat: add video dashboard config bar, hardware chart, KPI cards and t…
edwingao28 Sep 20, 2026
490e4be
feat: make the cross-hardware dashboard the /video landing view
edwingao28 Sep 20, 2026
cebb645
test: cover the video dashboard end to end with retained fixtures
edwingao28 Sep 20, 2026
2cab668
docs: explain the VideoGenX dashboard data path, axes and C1 default
edwingao28 Sep 20, 2026
5c50146
fix: keep video KPI swatches hydration-safe and show placeholders whi…
edwingao28 Sep 20, 2026
d189dce
fix: align video dashboard Chinese labels with existing tier and metr…
edwingao28 Sep 20, 2026
5d45191
feat: key video cells by deployment and carry the replica count
edwingao28 Sep 20, 2026
63a3d77
feat: draw per-hardware Pareto frontiers over deployments, never thro…
edwingao28 Sep 20, 2026
8532946
feat: add H3 API price reference and API-priced video metrics
edwingao28 Sep 20, 2026
8751d9e
feat: carry the API price in the video dashboard URL state
edwingao28 Sep 20, 2026
92e9130
feat: add API price reference control and revenue/profit views to the…
edwingao28 Sep 20, 2026
c7273ac
test: cover the API price reference in video component specs
edwingao28 Sep 20, 2026
77e6257
fix: cap the video chart y axis at break-even when every hardware los…
edwingao28 Sep 20, 2026
5bd497b
feat: add video compare metric deltas and case pairing
edwingao28 Sep 20, 2026
dccdf03
feat: add compare URL selection for baseline, candidate and case
edwingao28 Sep 20, 2026
1a5116c
feat: add arena-style VideoCompare panel with on-demand clips
edwingao28 Sep 20, 2026
8a0a092
fix: skip malformed sources when resolving compare clips
edwingao28 Sep 20, 2026
66099c8
feat: derive compute-bound evidence from measured video cells
edwingao28 Sep 20, 2026
e49c63b
feat: add VideoEvidence compute-bound panel
edwingao28 Sep 20, 2026
d7cfae9
test: cover the video evidence panel with the retained fixture
edwingao28 Sep 20, 2026
3172afa
fix: bound the power meter ARIA value to its declared 0–100 range
edwingao28 Sep 20, 2026
1496365
docs: note VideoEvidence reads only the basis from its metric options
edwingao28 Sep 20, 2026
67e91ce
refactor: pick compare and evidence cells by deployment layout
edwingao28 Sep 20, 2026
e7b4be9
feat: wire the API reference, Compare and Evidence panels into the vi…
edwingao28 Sep 20, 2026
51a13cc
docs: describe the API reference, Compare and Evidence sections of th…
edwingao28 Sep 20, 2026
d0a5982
fix: keep the API price legible and describe Compare cells as lead de…
edwingao28 Sep 20, 2026
19cd5f2
fix: polish the Chinese copy of the video dashboard panels
edwingao28 Sep 20, 2026
9e507dd
feat: make the video Compare panel an arena-style side-by-side player
edwingao28 Sep 20, 2026
a0ee769
refactor: trim the video dashboard to the cost and compute-bound claims
edwingao28 Sep 20, 2026
7bbdf34
Merge branch 'master' into feat/videogenx-hardware-dashboard
edwingao28 Sep 20, 2026
cfa834a
refactor: remove the legacy /video viewer, trade-off chart and CI-run…
edwingao28 Sep 20, 2026
ba7fd84
test: drop the retired GPU-basis option from the video deployment tests
edwingao28 Sep 20, 2026
25ee536
fix: read one frozen workload and skip under-sampled cells on the vid…
edwingao28 Sep 21, 2026
3d47c3a
chore: merge master and keep the legacy video viewer removed
edwingao28 Sep 22, 2026
a558876
fix: keep the video dashboard URL write out of the state updater
edwingao28 Sep 24, 2026
0a63ac4
feat: align the video dashboard chrome with the inference tab
edwingao28 Sep 24, 2026
443bc6c
fix: source the H3 API reference price from the official pay-as-you-g…
edwingao28 Sep 24, 2026
8cf35fd
fix: preserve VideoGenX comparison and API contracts
edwingao28 Sep 26, 2026
debc381
chore: merge master into VideoGenX comparison branch
edwingao28 Sep 27, 2026
ac73782
fix: keep video concurrency evidence in its shared cohort
edwingao28 Sep 27, 2026
06a61d3
feat: add video deployment frontiers and Wan serving evidence
edwingao28 Oct 5, 2026
6bcddb0
feat: add read-only H3 performance history
edwingao28 Sep 15, 2026
25299d4
fix: open published H3 results from the video tab
edwingao28 Sep 16, 2026
076a6ee
feat: map H3 artifact hardware to HW_REGISTRY keys and TCO tiers
edwingao28 Sep 20, 2026
c66febc
feat: add video benchmark metric registry with TCO tiers and GPU basis
edwingao28 Sep 20, 2026
8d7974f
feat: add direction-aware Pareto frontier for video hardware points
edwingao28 Sep 20, 2026
9b38e5c
feat: add v_ URL state for the video hardware dashboard
edwingao28 Sep 20, 2026
fd056ac
feat: project GPU counts, power and server layout per published video…
edwingao28 Sep 20, 2026
d5743da
feat: flatten published video history into chart points
edwingao28 Sep 20, 2026
bed9d02
test: rebuild video history fixture with per-cell GPU, power and layo…
edwingao28 Sep 20, 2026
b8a8502
feat: add video dashboard config bar, hardware chart, KPI cards and t…
edwingao28 Sep 20, 2026
e74d5b8
feat: make the cross-hardware dashboard the /video landing view
edwingao28 Sep 20, 2026
8f02188
test: cover the video dashboard end to end with retained fixtures
edwingao28 Sep 20, 2026
30b9cd5
docs: explain the VideoGenX dashboard data path, axes and C1 default
edwingao28 Sep 20, 2026
2ae4189
fix: keep video KPI swatches hydration-safe and show placeholders whi…
edwingao28 Sep 20, 2026
70bc228
fix: align video dashboard Chinese labels with existing tier and metr…
edwingao28 Sep 20, 2026
2a671b9
feat: key video cells by deployment and carry the replica count
edwingao28 Sep 20, 2026
9d542cc
feat: draw per-hardware Pareto frontiers over deployments, never thro…
edwingao28 Sep 20, 2026
f37161a
feat: add H3 API price reference and API-priced video metrics
edwingao28 Sep 20, 2026
b54404f
feat: carry the API price in the video dashboard URL state
edwingao28 Sep 20, 2026
da9d2e1
feat: add API price reference control and revenue/profit views to the…
edwingao28 Sep 20, 2026
2b63f52
test: cover the API price reference in video component specs
edwingao28 Sep 20, 2026
eebadd6
fix: cap the video chart y axis at break-even when every hardware los…
edwingao28 Sep 20, 2026
13e4adc
feat: add video compare metric deltas and case pairing
edwingao28 Sep 20, 2026
af453d2
feat: add compare URL selection for baseline, candidate and case
edwingao28 Sep 20, 2026
549101f
feat: add arena-style VideoCompare panel with on-demand clips
edwingao28 Sep 20, 2026
3a85cbb
fix: skip malformed sources when resolving compare clips
edwingao28 Sep 20, 2026
0411e69
feat: derive compute-bound evidence from measured video cells
edwingao28 Sep 20, 2026
934946c
feat: add VideoEvidence compute-bound panel
edwingao28 Sep 20, 2026
78782b2
test: cover the video evidence panel with the retained fixture
edwingao28 Sep 20, 2026
4e7f2c5
fix: bound the power meter ARIA value to its declared 0–100 range
edwingao28 Sep 20, 2026
f8b492f
docs: note VideoEvidence reads only the basis from its metric options
edwingao28 Sep 20, 2026
e6308e4
refactor: pick compare and evidence cells by deployment layout
edwingao28 Sep 20, 2026
85f5220
feat: wire the API reference, Compare and Evidence panels into the vi…
edwingao28 Sep 20, 2026
230e257
docs: describe the API reference, Compare and Evidence sections of th…
edwingao28 Sep 20, 2026
c4200d3
fix: keep the API price legible and describe Compare cells as lead de…
edwingao28 Sep 20, 2026
e39312e
fix: polish the Chinese copy of the video dashboard panels
edwingao28 Sep 20, 2026
2da595b
feat: make the video Compare panel an arena-style side-by-side player
edwingao28 Sep 20, 2026
e635a59
refactor: trim the video dashboard to the cost and compute-bound claims
edwingao28 Sep 20, 2026
9bc6baa
refactor: remove the legacy /video viewer, trade-off chart and CI-run…
edwingao28 Sep 20, 2026
cf9b02c
test: drop the retired GPU-basis option from the video deployment tests
edwingao28 Sep 20, 2026
f1dcade
fix: read one frozen workload and skip under-sampled cells on the vid…
edwingao28 Sep 21, 2026
50700f7
fix: keep the video dashboard URL write out of the state updater
edwingao28 Sep 24, 2026
50dae16
feat: align the video dashboard chrome with the inference tab
edwingao28 Sep 24, 2026
f81ab20
fix: source the H3 API reference price from the official pay-as-you-g…
edwingao28 Sep 24, 2026
d01b7e6
fix: preserve VideoGenX comparison and API contracts
edwingao28 Sep 26, 2026
ab386e8
fix: keep video concurrency evidence in its shared cohort
edwingao28 Sep 27, 2026
f0c8112
chore: adapt the video dashboard to master's lib layout and timing ba…
edwingao28 Oct 17, 2026
3b6e401
feat: accept native srt-slurm H3 bundles in the video dashboard
edwingao28 Oct 17, 2026
0520af4
feat: retain the native H200 run in the video fixture and read hardwa…
edwingao28 Oct 17, 2026
c9a1087
fix: read published clips in Compare and keep its URL-restored sides …
edwingao28 Oct 17, 2026
6bcd1b8
docs: state that compare sides stay distinct in the views API reference
edwingao28 Oct 17, 2026
d18d5c2
fix: start every Compare change from the pair on screen
edwingao28 Oct 17, 2026
dfae3ac
chore: sync VideoGenX delivery with master
edwingao28 Oct 18, 2026
cf1540a
fix: publish every native video sweep artifact
edwingao28 Oct 18, 2026
cb2f31d
fix: publish every native video sweep artifact
edwingao28 Oct 18, 2026
080b759
feat: preserve VideoGenX model and engine identities
edwingao28 Oct 18, 2026
9f2a282
chore: sync reviewed VideoGenX publication fixes
edwingao28 Oct 18, 2026
e484065
test: cover selected native video writer artifacts
edwingao28 Oct 18, 2026
cdb2968
fix: preserve unknown native GPU allocation
edwingao28 Oct 18, 2026
16dde84
fix: inherit native GPU allocation evidence rules
edwingao28 Oct 18, 2026
61113f8
fix: refresh VideoGenX skills integrity
edwingao28 Oct 18, 2026
c6d3425
fix: inherit refreshed VideoGenX skills integrity
edwingao28 Oct 18, 2026
File filter

Filter by extension

Filter by extension


Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
4 changes: 3 additions & 1 deletion .github/workflows/h3-media.yml
Original file line number Diff line number Diff line change
@@ -1,6 +1,8 @@
name: 'Publish H3 media'

on:
repository_dispatch:
types: [publish-video-results]
schedule:
- cron: '*/15 * * * *'
workflow_dispatch:
Expand Down Expand Up @@ -31,5 +33,5 @@ jobs:
bun-version-file: package.json
- name: Persist completed H3 artifacts
env:
H3_RUN_ID: ${{ inputs.run_id }}
H3_RUN_ID: ${{ inputs.run_id || github.event.client_payload.source_run_id }}
run: bun packages/app/scripts/warm-video-results.ts
4 changes: 4 additions & 0 deletions .prettierignore
Original file line number Diff line number Diff line change
Expand Up @@ -11,3 +11,7 @@ packages/app/cypress/fixtures/api/collectivex-runs.json
# The power-model parity fixture is upstream Python output, regenerated wholesale by
# packages/app/scripts/generate-power-model-parity.py; pretty-printed it is ~5x larger.
packages/app/src/lib/power/power-model.parity.json
# The native H3 bundle fixture is a checksum-sealed copy of the InferenceX producer's
# output: SHA256SUMS and the manifest bind every byte, so reformatting breaks the seal.
packages/app/src/components/video-benchmark/fixtures/native-serving-bundle/
packages/app/src/components/video-benchmark/fixtures/native-startup-failure/
75 changes: 73 additions & 2 deletions docs/dashboard-readonly-views.md
Original file line number Diff line number Diff line change
Expand Up @@ -75,7 +75,7 @@ combinations, not the full Cartesian product of all possible filter values.
| `reliability` | `asOf`, `format`, `gpus`, `range` |
| `submissions` | `direction`, `limit`, `lines`, `mode`, `offset`, `onChangeOnly`, `search`, `sort` |
| `ubenchx` | `gpu`, `test` |
| `video` | `artifact`, `cell`, `compare`, `costs`, `gpuBasis`, `page`, `phase`, `run`, `selected`, `slot`, `source`, `view`, `workload`, `xAxis`, `yAxis` |
| `video` | `artifact`, `cell`, `compare`, `costs`, `gpuBasis`, `page`, `phase`, `run`, `selected`, `slot`, `source`, `v_x`, `v_y`, `v_tier`, `v_view`, `v_optimal`, `v_api`, `v_hidden`, `v_base`, `v_cand`, `v_case`, `view`, `workload`, `xAxis`, `yAxis` |

Cache reuse returns `data.recipes` and the resolved `params.recipe`. Pass a returned key as `recipe` (the UI uses `c_recipe`) to select the same TP/EP/DP-attention, worker, GPU, speculation, offload and fingerprint combination. Omitted or stale keys use the shared dashboard default. Runs use that recipe if available, otherwise their own best-covered recipe; inspect each bar's source row for its identity. Overlay-only configurations expose their own recipe choices. Layout orientation and label placement are presentation-only controls.

Expand Down Expand Up @@ -120,9 +120,74 @@ alignment changes no API parameters, response schemas, or stored model identitie
- GPU metrics: shared line/correlation transforms and full-record per-GPU statistics
computed from the stored samples, or from the samples of live artifacts. File/host
identity and missing-versus-zero semantics persist.
- Video: checksum-verified stored bundles, serving/fidelity selectors and tradeoffs.
- Video: published history pages → videoPoints → dashboardCells → shared metrics,
deployment Pareto plot/table, lead-cell KPIs, Compare and measured evidence.
Legacy checksum-verified bundle serving/fidelity and tradeoff reads remain available.
- Overview/rankings/compare: existing discovery-page assembly and scenario helpers.

## VideoGenX dashboard contract

`GET /api/v1/views/video` defaults to the current `/video` dashboard. The shared
`VIDEO_HISTORY_MAX_PAGES` bounds both UI and API to the first five published
history pages; responses report `coverage` with `pagesRead`, `maxPages`,
`nextPage` and `truncated`. This is a publication-ordered window, not an execution
date range or a promise to enumerate the entire catalog. Failures remain failures;
missing fields remain null. Source run/artifact IDs, seals and distinct execution
and publication times remain in `provenance`; raw documents, private asset URLs
and credentials are excluded from the dashboard projection.

| UI control or surface | Public projection / scope |
| ------------------------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| X/Y, cost tier, optimal | `v_x`, `v_y`, `v_quality`, `v_qmin`, `v_tier`, `v_optimal`; same metric registry, participating-GPU cost basis and per-hardware frontier helpers. Queueing cells never enter `plot`/`rows`. |
| Quality view/filter | `v_y=quality`, `v_quality` (seven original human dimensions) and optional `v_qmin` (0–4). Require explicit ordinal_0_to_4 scale and calibrated pass for all seven critical dimensions, complete assessed coverage matching the point, protocol/rubric/evaluator hashes and frozen calibration provenance before Pareto selection. The reader filter cannot relax the frozen threshold. Chart/table/CSV/API agree; raw judged-but-unqualified scores remain distinct from derived eligible values. |
| Hardware visibility | `v_hidden`, sorted/deduplicated known roster keys; affects plot/table only, matching UI. KPI, Compare and Evidence retain all selected-workload cells. |
| API reference price | `v_api`, positive USD/video-second rounded to four decimals; only `apiPricePerVideo`, never TCO or revenue. Invalid values use the UI reference. |
| Chart/table | `v_view`; echoed as `params.displayView`, no numerical difference. Color, zoom and playback remain presentation state. |
| Workload/newest cell | `videoPoints` → `dashboardCells`; same canonical identity, sample floor and publication precedence. Return `workload`, `otherWorkloads`, cells and source provenance. |
| KPI cards | `kpis`: roster entries, shared lead cell and all metric values; missing hardware keeps null point/metrics. |
| Compare selectors/deltas | `v_base`, `v_cand`, `v_case`; shared selection defaults and `compareMetrics`. Default mode returns deltas without loading media. |
| Case media metadata | `view=compare` adds at most two published-only `format=published` reads (deduplicated by run/artifact), `compareSide`/`pairCases`, bounded `resolvedCaseIndex`, matched/unmatched counts and selected records. Only bundle-relative media paths are exposed. Missing media, missing baseline/candidate and read errors have separate statuses; `ready` means readable artifacts and can still have zero pairs. |
| Serving outcomes | `serving` retains every loaded planned cell before chart/workload/hardware/quality filters, including zero-attempt failures. Attempted failures and not-started slots remain separate; timeouts are a failure subset. Unknown populations remain null, and `qualitySloGoodput` is null without per-request quality/deadline evidence. |
| Evidence | Shared power/plateau/scaling/facts functions. Descriptive measurements; no quality qualification or causal bottleneck inference. |
| Legacy run browser | Explicit `view=discovery&page=…`; supplying `run`/`artifact` retains results and optional tradeoff selectors. Mixing legacy selectors and `v_*` returns 400. |

Dashboard state values use the UI parser and its fallback defaults; unknown or
repeated parameter names return 400. Legacy source status handling and 204 behavior
remain unchanged. Chart history read failures preserve the sanitized upstream status;
optional Compare media failure returns `comparison.cases.status=unavailable` without
erasing independently available chart data. The route and generated OpenAPI remain
no-store.

A measured frontier is grouped by explicit workload/generation identity and hardware;
quality mode additionally requires the same contract, dimension, evaluator and
calibration cohort. Hardware-health failures and unknown request capacity are
excluded from ordinary performance rankings. A single point remains a point;
lines join measured eligible points and do not extrapolate. `quality: null`
means unjudged, and quality modes return an honest empty selection until calibrated
evidence exists. Integrity checks and numerical fidelity are separate evidence.

Regression coverage exercises numerical API/UI parity across default and changed
axes/tier/optimal/price/visibility, the shared five-page bound, legacy reads,
case-index clamping, absent/failed media and private-URL exclusion. This does not
claim live production API/UI equivalence or full-history completeness.

### 中文说明

`v_y=quality` 展示单项人工质量评分与延迟,`v_quality` 选择七个独立维度之一,
`v_qmin` 在 Pareto 计算前按 0–4 级阈值筛选。必须具备完整样本覆盖、已校准的通过
判定及协议、rubric、evaluator 和校准来源记录;读者阈值不能放宽冻结的规则。
未评价或未校准数据保留为空,图表、表格、CSV 和 API 选择一致。原始已评但未获
资格的分数可作为证据读取,不能进入质量前沿。硬件健康失败及未知 batch 容量不会
进入普通性能排名;不同 workload、生成配置或质量 evaluator 不合并连线。

视频公开接口默认复现当前仪表板:最多读取前五页已发布历史,复用相同的 workload、
样本门槛、部署和数值计算,明确返回覆盖范围与截断状态。`v_hidden` 只隐藏图表和表格
中的硬件,KPI、Compare 和 Evidence 保持原范围;`v_api` 只是视频秒参考价,不是
TCO 或收益。默认不加载媒体;`view=compare` 才读取最多两份已发布媒体产物并返回
匹配用例、实际索引和未配对数量,不暴露 asset URL 或原始数据包。`view=discovery`
及原有 run/artifact 查询仍可使用;不能与 `v_*` 混用。未发布、无可比较双方和读取失败
分别标记,不能解释为成功的空测量。该接口不承诺完整历史、质量验收或瓶颈因果结论。

## Other public surfaces

Model, chip and pair discovery pages reuse overview/rankings/compare calculations.
Expand Down Expand Up @@ -455,3 +520,9 @@ combinations from pending qualification. Footnotes contain both `en` and `zh`.
CollectiveX 视图另返回 `knownSupport` 和 `knownSupportFootnotes`,与网站共用人工维护
的数据源,不受所选运行影响。状态值为 `works`、`broken` 和 `na`;请结合注释区分
不可用与待验证的配置。注释同时包含 `en` 和 `zh`。

### Video engine identity

Published history exposes `producer` when the manifest, matrix and execution records agree on the selected `plan_sha256`, model revision and `engine: {name, revision}`. Legacy sources retain null identity. The dashboard and API expose `releaseCoverage` for H3/Wan × SGLang/vLLM-Omni across the loaded published pages, including missing, failed, history-only and sample-floor-met states. This bounded coverage is not a full-catalog or quality qualification claim. Native four-clip runs remain History-only; the chart, KPI and Compare sample floor stays ten.

Compare keeps distinct engine revisions on the same GPU. `v_base` and `v_cand` accept `hardware:engine:revision`; hardware-only links remain valid. Technical deltas require the same sealed model revision, selected plan and generation settings. Unknown quality does not suppress technical observations, but does not authorize quality-normalized claims. CSV rows include `model`, `model_revision`, `engine_revision` and `plan_sha256`. Compare reads `format=published` only.
2 changes: 2 additions & 0 deletions docs/index.md
Original file line number Diff line number Diff line change
Expand Up @@ -22,6 +22,7 @@ Design rationale and non-obvious conventions. See [CLAUDE.md](../CLAUDE.md) for

- [Architecture](./architecture.md) — Why client-first, route navigation, URL state, provider nesting, server-side caching (unstable_cache + blob), in-memory client cache, color system, analytics enforcement
- [D3 Charts](./d3-charts.md) — Why 4 effects, in-place mutation, refs for zoom, rAF throttling, HTML tooltips, Pareto directions, gradient labels
- [Video History](./video-history.md) — Read-only published H3 catalog, metric boundaries, sharing, and local retained-data replay
- [Data Pipeline](./data-pipeline.md) — DB schema reasoning, ETL design, transform pipeline, spline method choice, normalizer resolution order (model/GPU/framework)
- [Pitfalls](./pitfalls.md) — Failure modes: token type consistency, schema evolution, empty objects, zoom loss, stale closures, disaggregated metrics, negative splines, date stamping, ref stability, cost inheritance
- [GPU Specs](./gpu-specs.md) — Unit conventions, topology invariants, SVG layout rationale, hardware gotchas
Expand All @@ -34,6 +35,7 @@ Design rationale and non-obvious conventions. See [CLAUDE.md](../CLAUDE.md) for
- [State Ownership](./state-ownership.md) — Which context owns which state, availability filtering cascade, comparison date mechanics, URL param sync
- [Blog](./blog.md) — MDX content system, SEO features (OG images, RSS, llms.txt, JSON-LD), TOC sidebar, reading progress, heading links, analytics events
- [CollectiveX](./collectivex.md) — lazy ingest-on-read pipeline (separate Neon DB as a durable GitHub-artifact cache), tombstoned deletes, raw-docs storage + shared reader
- [VideoGenX dashboard](./videogenx-dashboard.md) — `/video` cross-hardware chart: published-history projection → points → metrics, TCO tiers per participating GPU vs. the API list price, deployment cells (queued cells never plotted), arena Compare, performance evidence, unavailable-hardware roster, null rules
- [Chinese Pages (/zh)](./i18n.md) — Why hand-authored /zh pages instead of an i18n framework, hreflang pairing, blog translation pairing, html lang workaround, CJK reading time/slugs
- [Chinese Copy Editorial Guide](./chinese-copy.md) — Audience, surface-specific register, context-aware terminology, rewrite workflow, manual-review process, and PR checklist

Expand Down
Loading
Loading